RFDiffusion
1 Introduction
RFdiffusion1,2 is an open source method for structure generation, with or without conditional information (a motif, target etc). In a manner analogous to networks which produce images from user-specified inputs, RFdiffusion enables the design of diverse, complex, functional proteins from simple molecular specifications.
Can do
- Motif Scaffolding.
- Unconditional protein generation.
- Symmetric unconditional generation.
- Symmetric motif scaffolding.
- Binder design.
- Design diversification ("partial diffusion", sampling around a design).
Figure 1. The Process summary diagram of RFDiffusion.
2 Parameters
General
| Name | Explanation |
|---|---|
| task type | The type of task. |
| input pdb | The PDB file of three-dimensional structure. |
| target patch | Specify target objects by PDb chain and position. E.g. A1-150. |
| num designs | The number of designs. |
Motif Scaffolding
| Name | Explanation |
|---|---|
| pre length | The length of the amino acids generated before target objects. E.g. 10-40. |
| post length | The length of the amino acids generated after target objects. E.g. 60-60. |
Binder Design
| Name | Explanation |
|---|---|
| hotspot res | The hotspot residue in PDB. E.g. [A59,A83,A91]. |
| output length | The length of the amino acids generated. E.g. 70-100. |
3 Results Explanation
The .pdb file. This is the final prediction out of the model. Note that every designed residue is output as a glycine (as we only designed the backbone), and no sidechains are output. This is because, even though RFdiffusion conditions on sidechains in an input motif, there is no loss applied to these predictions, so they can't strictly be trusted.
4 Reference
[1] Watson, J.L., Juergens, D., Bennett, N.R. et al. De novo design of protein structure and function with RFdiffusion. Nature 620, 1089–1100 (2023). https://doi.org/10.1038/s41586-023-06415-8
[2] Watson JL, et al. Broadly applicable and accurate protein design by integrating structure prediction networks and diffusion generative models. https://doi.org/10.1101/2022.12.09.519842